The problem
The sales team requested a significant video production budget because "reps aren't using video." Management was prepared to fund it. But the real issue wasn't content scarcity—it was content chaos.
What was actually happening
- 3,400+ videos and PDFs existed in the repository.
- Video search returned irrelevant results (searching "demo" returned training videos, podcasts, and old unrelated content).
- Navigation required knowing the old taxonomy—product line folders, then legacy categories, then file names.
- New content was impossible to find because metadata was inconsistent.
- Reps didn't know what existed, so they asked the org to make more instead of looking for what was already there.
Discovery & diagnosis
User research
- Conducted 15 interviews with reps across regions and experience levels.
- Shadowed 8 sales calls to understand when and how reps needed video content.
- Analyzed search logs: 34% of searches returned zero results, and when they did, most results were irrelevant.
- Reviewed repository content: found videos on the exact topics reps were requesting, just undiscoverable.
The insight
Reps needed videos for 4 key moments: discovery calls, objection handling, demo walkthroughs, and closing. But the repository was organized by content creator and date, not by use case. A rep doing a discovery call would have to know it existed and know where to look.
Approach
Reframing the investment
I argued: before spending $150K on new production, spend $20K on organizing and improving discovery of what exists. If that doesn't work, then we invest in production. Leadership agreed to a 60-day pilot.
Engineering & UX collaboration
Collaborated with engineering and UX teams to define product requirements, worked through technical constraints and solution trade-offs, and iterated on design based on user feedback and technical feasibility. When Engineering initially pushed for legacy infrastructure, I made the case for a new tagging system based on user research data—and we shipped it.
The redesign
- New tagging system: Instead of folder hierarchy, every video got tagged by use case, product, and seniority level. A rep could find "discovery call videos for new SDRs" or "premium tier objection handling."
- Improved search: Integrated AI-powered search that understood sales intent, not just keywords.
- Metadata cleanup: Standardized titles, descriptions, and metadata across 3,400+ assets.
- Promoted content: Created "curated playlists" for common sales moments (first call, handling objections, demo).
- New UI: Replaced the file-browser experience with a modern card-based grid discoverable by use case.
- Rating system: Let reps rate videos so the most useful content floats to the top.
Execution roadmap
- Month 1: Data audit, tagging system design, and quick-win content tagging (300 top videos).
- Month 1-2: Full repository tagging (all 3,400 assets).
- Month 2: UI redesign and search improvement.
- Month 2-3: Launch, monitoring, iterate on taxonomy based on search behavior.
Results
Drove 691% increase in video views within one month of launch. The platform became the primary video delivery system for 3,000+ sales reps across Meta's global sales org.
Engagement metrics
- 691% increase in video views in Q1 post-launch (compared to Q4 baseline).
- 57% increase in PDF downloads.
- Average session time in repository increased 3.4x.
- Search results with zero hits dropped from 34% to 4%.
- Most-viewed content shifted: the top 10 most-watched videos had 0 overlap with pre-launch top 10 (indicating the new taxonomy was surfacing previously hidden content).
Business impact
- Production budget saved: The $150K request was rejected. The $20K investment in discovery returned ROI.
- Content ROI improved dramatically: Existing content that cost $0 in Q4 now drives engagement. The org realized they were sitting on a goldmine.
- Ramp time improved: New reps found onboarding videos 84% faster, improving time-to-first-call.
- Precedent set: This changed how the org thinks about content strategy. Now the question is "Can we make existing content more discoverable?" before "Should we make new content?"
Key learnings
1. Organization beats production. An organized library of old content beats a disorganized library of new content. Reps use what they can find.
2. Listen to the problem, not the ask. The team asked for more video. The real problem was "I can't find what I need." Those aren't the same problem.
3. Metadata matters. How you tag and categorize content determines if it gets used. Spend time on taxonomy.
4. Measure what indicates real value. Vanity metrics: "We made more videos." Real metrics: "Reps found what they needed faster and it showed up in their sales performance."